For Walmart, it’s a priority to quickly stock store shelves with the freshest food possible.
In doing so, the nation’s largest grocery-store retailer can prevent food waste from eating into its profits, even in an industry known for thin margins.
Yasemin Gunay, a managing director at Boston Consulting Group, told Business Insider that retailers and restaurants have historically planned their inventory levels based on their internal data, mostly based on past sales trend data.
However, advancements in AI algorithms have allowed companies like Walmart to consider external factors, like changes to the weather or major sporting events. By more accurately anticipating demand, retailers can be more strategic with their pricing strategies, employing more targeted promotions based on their stock levels and local sales trends.
“There’s a lot of data in retail and you can monetize it if you can actually take action on it,” said Gunay.
Inside Walmart’s AI-assisted manufacturing facilities
Over the past four years, Walmart has constructed and opened several high-tech grocery distribution centers in states like California, South Carolina, and Texas. Each facility has automated robotics and AI systems that support workers’ ability to process fresh vegetables and perishables more quickly than at traditional facilities without such technologies.
One of those facilities, in Wellford, South Carolina, is a 725,000-square-foot distribution center that autonomously accepts deliveries from vendors and ships out those goods to 180 Walmart stores across five states.
James Bright, the general manager of the Wellford facility, told Business Insider that the distribution center is 98% automated. The site relies on robotics to review and process deliveries of frozen food, dairy goods, deli meats, and other perishables from farmers and food suppliers. The goods are then stored in temperature-controlled environments to keep everything fresh. Another set of autonomous machinery sorts those goods into pallets ready for store delivery.
Bright said that AI helps enable the process of creating the “perfect pallet.” AI algorithms help to determine the optimal pallet placement for each food item, including how many gallons of milk or packs of lettuce are needed in each container, the ideal configuration to fit as many items in a pallet as possible, and take into consideration weight to ensure lighter items like eggs aren’t crushed.
Algorithms can also triangulate data from individual stores’ layouts, placing the items in an order that makes it easier for associates to remove the goods and stock them on the shelves, Bright said.
He added that Walmart also considers feedback from associates. If employees request adjustments, the retailer can tweak the algorithms to make them more accurate.
“A lot of use of AI is in the background as we continue to automate our supply chain,” said Bright.
According to a company blog post, Walmart uses predictive AI in Costa Rica to map out the best delivery routes for pineapples and root vegetables. In Mexico, Walmart has launched an AI-enabled inventory tool that tracks when goods may be at risk of being overstocked and automatically reroutes the food to other stores with less supply.
How AI can sharpen inventory planning
Indira Uppuluri, Walmart’s senior vice president of supply chain technology, said AI is helping the company forecast demand beyond the immediate future. This can help lower costs, because Walmart incurs more costs by housing excess inventory. “You want to reduce the buffer inventory,” said Uppuluri.
Walmart also uses AI to respond to demand fluctuations, according to Parvez Musani, the senior vice president of stores and online pickup and delivery technology. For example, AI models can be trained to send extra shovels to stores in Arizona for an upcoming snowstorm, but won’t increase the stock levels for Walmart’s Maine stores, where snow is far more prevalent.
AI also helps Walmart plan ideal transportation routes for truck drivers delivering to the store and drivers who drop off express-order products that are frequently shipped to shoppers’ homes within an hour, Uppuluri said.
“There are multiple places where today we have generative AI in production being leveraged by our associates,” said Uppuluri. “It’s not just a proof of concept.”
